A Three Level AI Competency Framework for Inclusive AI Adoption

Canada is a recognized world leader in artificial intelligence (AI) research but has been slower to spread AI and digital tools across its economy, especially among smaller firms. The federal AI for All strategy now puts adoption, productivity, skills and shared prosperity at the centre of the national approach. The question is no longer only whether people and organizations can access AI, but whether they have the capacity to benefit from it.

Discussion about AI skills tends to fall into one of two extremes: the technical elements of building and maintaining AI systems on the one hand, and the broad literacy that helps people understand what AI is on the other. Both matter, and neither is enough for adoption. AI is reshaping jobs by changing the everyday tasks people do, the tools they use and the outputs they must check. Most exposed jobs are more likely to be reorganized rather than eliminated, and the skills they i demand are shifting quickly, with one analysis The Diversity Institute AI finding junior roles up to seven times more likely to require skills traditionally associated with senior roles, such as judgment and decision-making. The same tool can call for very different competencies depending on the task, the information involved and who is affected; drafting a marketing email with an AI assistant is markedly different from using AI in hiring, health care or financial decisionmaking. Competency therefore reaches well beyond either coding or prompting to include judgment, verification, communication and domain knowledge.

Two realities underpin all of this. First, responsible AI use is a workplace competency, not only a governance concern. Many risks arise through ordinary use, such as “shadow AI,” where employees paste sensitive information into public tools; in one survey, 46% admitted doing so. Protecting information, checking outputs and knowing when to escalate are practical, daily skills. Second, AI competency is a matter of equity. Access, workplace exposure, training and professional networks are distributed unevenly across gender, race, age, disability, geography and Indigeneity, from connectivity gaps affecting rural areas and First Nations reserves to lower workplace AI use among older workers. Women face compounded barriers: their occupations are nearly twice as exposed to generative AI as men’s (29% versus 16%), yet women make up only about 30% of the global AI workforce, leaving them under more pressure to adapt while under-represented among those shaping the technology. Who develops and who benefits from AI competency is therefore not a neutral question.

Naming the right capabilities is only the first step, because people and organizations still need the conditions to develop and use them. Competency grows when learning is tied to real tasks, offers genuine chances to practise, and is backed by the tools, time and organizational support needed to apply it. It also depends on addressing the barriers that shape who can participate and benefit; equity must shape outreach, delivery and support from the outset rather than being added later. SMEs face particular constraints of time, budget and formal training capacity, making organizational support and access to trusted intermediaries important for turning learning into lasting organizational capability. Success should also be measured by what people can do and whether those competencies are being applied, not only by enrolment or course completion.